American University in Cairo
AUC Knowledge Fountain
Papers, Posters, and Presentations
2011
The evolution of Islamic arches till 1250
Ehsan Abushadi
Follow this and additional works at: https://fount.aucegypt.edu/studenttxt
Part of the Near Eastern Languages and Societies Commons
Recommended Citation
Abushadi, Ehsan, "The evolution of Islamic arches till 1250" (2011). Papers, Posters, and Presentations.
37.
https://fount.aucegypt.edu/studenttxt/37
This Other is brought to you for free and open access by AUC Knowledge Fountain. It has been accepted for
inclusion in Papers, Posters, and Presentations by an authorized administrator of AUC Knowledge Fountain. For
more information, please contact fountadmin@aucegypt.edu.
Abushadi
1
Ehsan Abushadi
ARIC 270
O’Kane
Draft - Term Paper
April 27, 2011
The Evolution of Islamic Arches till 1250
Arches are a common feature of early architecture; they were an effective architectonic
form originally used to support a building or roof, while creating a sense of vastness. It rapidly
integrated into the architecture that developed within Dar al-Islam. Their design and purpose was
simple, but as time progressed through the Islamic Empire, with dynasties replacing the previous
ones, arches started to develop in shape and decoration, becoming increasingly less simple
through the centuries. The development of arches was not monotonous or homogenous
throughout Dar al-Islam, due to its large expanse each area developed differently, being
influenced by local architecture, context, artisans and masons.
If we accept the generally correct 19th century theory “that buildings could be dated by
the shape of their arching”1, then we can clearly say that each type or shape of arch is specific to
one period. Therefore, these period specific arches cannot be of random shapes or forms as they
must form some sort of smooth transition from one period to the next. In other words, there is an
evolution of arches.
Amongst the first arches to be used are the semicircular arches. We find examples of
these early semicircular arches at the Great Mosque of Damascus2, Syria, after the modification
by Al-Walid, 705-15; arches that might have been inspired by those that existed at the temenos
1
John Warren, “Creswell's Use of the Theory of Dating by the Acuteness of the Pointed Arches in Early Muslim
Architecture”, in Muqarnas, no 8 (1991), p. 59
2
K.A.C. Creswell, Early Muslim Architecture, New York: Hacker Art Books, 1979, vol. 1.1, p. 161
Abushadi
2
prior to the arrival of the Muslims to Damascus. The semicircular arch is also adopted from
Byzantine architecture3, which was prominent amongst Christian monuments in Syria. Other
examples of the semicircular arch can be found at the Umayyad palace of Khirbat al-Minya4,
705-15, Lake Tiberias in Israel/Palestine; the mosque of Jabal Says5, early 700s, south east of
Damascus; and the Umar Mosque at Bosra6.
Fig. 1 Separation of the origin to form a two centre point arch. Increase in separation of the two centre points results
in greater acuteness in the arch (after Warren).
The semicircular arch developed so that its centre or origin separated into two, creating
the two-pointed arch (see Fig. 1). The two-pointed arch makes its first appearances in the Islamic
world under the Umayyad dynasty with examples found at Qusayr Amra7, 712-15, and at the
caldarium of the bath at Hammam as-Sarakh8, 725-30, identifying Syria as the place of origin of
the pointed arch in Islamic architecture9. According to Creswell, there were no prior examples of
the pointed arch, which did not make an appearance in European architecture till round about the
start of the twelfth century; showing that this was an entirely Islamic development10. However,
as early as the Babylonian period, there have been examples of the pointed arch in the region of
3
John Warren, op.cit., p. 59
K.A.C. Creswell, Early Muslim Architecture, New York: Hacker Art Books, 1979, vol. 1.2, p. 387
5
Ibid. p. 474
6
Ibid. p. 545
7
Ibid. p. 391
8
Ibid. p. 442
9
Ibid. p. 443
10
K.A.C. Creswell, Early Muslim Architecture, vol. 1.2, op.cit., pp. 442
4
Abushadi
3
Khorsabad11. Creswell’s error is due to his acceptance of how Herzfeld “had avowed that all
Sasanian arches were rounded”12, without carrying out further investigation.
In the two centuries after the introduction of the two-pointed arch in Islamic architecture,
the form started developing to give the three, four and five-pointed arches. These designs were
produced by moving the arrangement of the centre points onto two horizontal layers with a
vertical difference (see Fig. 2). An increase in the number of centre points “postulates
progressive acuteness and complexity in the curves. Their visual qualities become a stylistic
feature which evolves and hence is a factor of time. It is also circumstantial and therefore
indicates both date and origin.”13 Furthermore, there are many variations of each of pointed or
multiple pointed arches, where they are more or less pointed or acute. In other words, the slope
of the pointed arch varies. This is achieved by increasing or decreasing the separation between
the centre points (see Fig. 1).14 Two examples of arches with varied acuteness are: at the Ibn
Tulun mosque, 870-9, where they are short pointed arches15, and at the Fatimid mosque of alHakim (al-Anwar), 990-1013, where they are high pointed arches16.
Fig. 2 Design of the three
pointed, left, and four pointed
arches, right (after Poole).
11
Camilla Edwards & David Edwards, “The Evolution of the Shouldered Arch in Medieval Islamic Architecture”,
in Architectural History, no 42 (1999), p. 71
12
John Warren, op.cit., p. 62
13
Ibid, pp. 59-60
14
K.A.C. Creswell, Early Muslim Architecture, vol. 1.2, op.cit., pp. 443-4
15
Doris Behrens-Abouseif, Islamic Architecture in Cairo: An introduction, Cairo: The American University in
Cairo Press (1989), p. 63
16
Idem.
Abushadi
4
According to Edwards & Edwards the origins of the tri-lobed arch can be considered to
lie in the pointed arch, with “cusping at the haunches”17. This possible development can be
pictured more clearly if we look at the squinch of the interior dome of Masjid-I Jami at
Ardistan18, 118019, where we see that within the typical squinch that has a pointed arch outline,
lies another arch cusped to form the tri-lobed arch (see Fig. 3). Therefore, it can be inferred that
Iranian and Mesopotamian domes have influenced the creation of the tri-lobed arch.20 Hence it is
not surprising that the tri-lobed arch started developing in the area of Iran and Iraq. Variations of
the tri-lobed arch include the exaggerated tri-lobed arch, which has an elongated apex (see Fig.
4.2A); the flattened tri-lobed, which has a flattened apex, or segment (see Fig. 4.2B); and the trilobed arch with notch shoulders21 (see Fig. 4.2D). In turn, the tri-lobed arched, particularly in a
squinch, influenced the development of muqarnas.
Fig. 3 Squinch with embedded tri-lobed arch at Masjid-I Jami, Ardistan (after
Edwards & Edwards).
17
Camilla Edwards & David Edwards, op.cit., p. 75
Ibid. p. 72
19
Arthur Pope, Persian Architecture, London: Oxford University Press (1969), p.58
20
Camilla Edwards & David Edwards, op.cit., p. 72
21
Ibid, p. 70
18
Abushadi
5
Fig. 4 Selection of corbelled shouldered arches, series 1, and selection of tri-lobed arches and a keel arch, series 2,
(after Edwards & Edwards).
However, Edwards & Edwards also acknowledge that the tri-lobed arch has also been
greatly influenced by the shouldered arch (see Fig. 4.1A-D). An example of the corbelled
shouldered arch can be found at Zafaraniyya22 (see Fig. 5). “By the early Saljuq period (c. 10401100) we see the corbelled shouldered arch almost invariably associated with the tri-lobed
shouldered arch in a variety of arrangements, part-structural and part-decorative.”23 It became
commonly used in the area of Iran and Iraq; no doubt it was also the reason why the tri-lobed
arch was also popular in the region. Nevertheless the corbelled shouldered arch continued to be
used in Seljuk architecture without the tri-lobe variant, such as in the Seljuk phase of the Isfahan
Friday Mosque.
22
23
Camilla Edwards & David Edwards, op.cit., p. 71
Ibid, p. 75
Abushadi
6
The origins of the keel arch (see Fig. 4.2C) also known as the triangular arch, not to be
confused with the quasi-four-centred arch or four-pointed arch (see Fig. 2), are not as obvious as
the previous examples of arches. The keel arch can be identified for its “straight haunches”, as
opposed to curved haunches of the pointed arch24. Edwards & Edwards suggest that the keel arch
is a development of the shouldered arch just like the tri-lobed arch, as a result of having included
it in their article concentrating on the evolution of the shouldered arch. However, round about the
period when the keel arch started to become common, the arts were flourishing within the
Islamic Empire, many things were being created, so it should not be ruled out that the keel arch
is an abstraction of the pointed arch. The keel arch became popular during the Fatimid and
Ayyubid25 dynasties, making appearances at Mashhad al-Juyushi (of Badr al-Gamali), 108526;
the mosque of the Vizier al-Salih Tala’i’ (see Fig. 6), 116027; and at the madrasa of al-Salih
Najm al-Din Ayyub, 124328. The foundation dates correlate to the theory of the evolution of the
arches, the keel arches being more complex to draw or design than those previously encountered.
Fig. 5 Corbelled shouldered arch at the caravanserai of Zafaraniyya (after Gye).
24
Camilla Edwards & David Edwards, op.cit., p. 70
Ibid, p. 84
26
Doris Behrens-Abouseif, op.cit., p. 67
27
Ibid, pp. 76-7
28
Ibid, p. 89
25
Abushadi
7
Fig. 6 Entrance façade, with keel arches, of the mosque of the Vizier
al-Salih Tala’i’ (Source: http://archnet.org/library/images/oneimage.jsp?location_id=4870&image_id=113448).
Similarly it is unclear where the origin of the horse-shoe (see Fig. 7) arch lies. The first
example of the horse-shoe arch in Islamic architecture is at the Great Mosque of Damascus29.
Creswell examined the topic and came across various theories pertaining to the origins of the
horse-shoe arch. He discovered that Dieulafoy believed it to have been derived from Persia
having found a strong correlation between the horse shoe arch and the arches at the palace of
Firuzabad, 226, in the way they were constructed30. On the other hand, Brown suggested that it
“was the legitimate offspring of the Romanesque”31 arch, while Gomez-Moreno and Rivoira
believed the horse-shoe arch to be derived from India32. However, the arches Gomez-Moreno
and Rivoira used as reference in India were mostly found in Chaitya Halls, they were cut from
stone rather than built. Such examples are the Lomas Rishi (see Fig. 8), 257 BC, or the temple at
Nasik, 50 BC33. Built examples include those found in Christian Mesopotamia, namely the
29
K.A.C. Creswell, Early Muslim Architecture, vol. 1.2, op.cit., p. 173
Ibid, p. 198
31
Ibid, pp. 198-9
32
Ibid, p. 199
33
Idem.
30
Abushadi
8
Baptistery of Mar Yaqub, 359, at Nasibin34; in Hama, North Syria there is the Halban doorway
dating from 54335; the doorway of Shaykh Ali Kasun in Syria36; at the apse of the church in
Dana, 48337; and at the apse of the church of the west monastery at Dayr Siman, 6th century38.
Fig. 7 Horse-shoe arch, left, and the stilted horse-shoe
arch, right (after Poole).
Fig. 8 Lomas Rishi (source:
http://en.wikipedia.org/wiki/File:CunninghamMaur
yan.jpg).
It is very likely that the horse-shoe arches in the Great Mosque of Damascus were
influenced by the churches in Syria. However, it can well be that Dieulafoy, Brown, GomezMoreno and Rivoira are all correct. If we assume that this is the case, then the horse-shoe arch
could have originated in India. Around the time of the construction of Lomas Rishi and the
temple at Nasik, the Parthian Empire spanned to the borders of India. The Parthian empire which
was quite influential could have hired Indian artisans or had Indian prisoners of war which could
34
K.A.C. Creswell, Early Muslim Architecture, vol. 1.2, op.cit., p. 199
Idem.
36
Ibid, p. 200
37
Idem.
38
Idem.
35
Abushadi
9
have brought the influence of the early horse-shoe arch into central Asia. This influence could
have been passed into the construction of the Firuzabad palace and the Christian Mesopotamian
baptistery. Having already shown influence in a Christian place of worship, it could have easily
influenced neighbouring churches such as those in Syria, which was under Byzantine rule by the
time the Parthian Empire had given way to the Sassanian Empire. Romanesque architecture is a
combination of Roman and Byzantine architecture; therefore, Romanesque architecture could
have picked up on the influence of Byzantine Syria. Therefore, if this is correct, then all four
theories as to the origin of the horse-shoe arch are correct.
However, this does not explain how it evolved within the Islamic Empire. There is a large
time period, approximately two centuries, between the construction of the Great Mosque of
Damascus and the construction of other monuments where the horse-shoe arch becomes popular,
mostly under the Aghlabid and Spanish Umayyad dynasties. In a way this raises the question of
whether the popular Islamic horse-shoe arch was a direct influence of a previous form or if it was
genuinely derived afresh by abstracting the semicircular arch. The semicircular arch could have
been abstracted to create the horse-shoe arch by slightly vertically, from the point of origin of the
circle, raising the circle that creates the semicircular outline form of the arch.
There is an obvious evolution in the arch when it comes to its shape, but it also evolves in
its purpose in a building. Often the evolution of the shape is affected by its structural purpose in
the building. Arches function in a structural manner by helping to support the weight of the
building above an opening, to prevent that section from collapsing due to a lack of support.
Therefore, an arch is placed across an opening to distribute the weight of the building above it so
that it can maintain its shape without collapsing. The stones or bricks of the arches do not fall as
they are cut, shaped or arranged in a manner that supports all the other stones or bricks,
Abushadi 10
preventing them from falling. Mortar is used to help reinforce this, keeping all units aligned so
that they do not shift weakening the arch as a whole.
“For an arch of a given rise and span and carrying a particular configuration of loads,
there is an ideal shape, called the “linear arch”.”39 The form or shape of the linear arch is
determined by the weight it carries so that “the linear arch represents the form which the real
arch should take to carry the particular configuration of loads with maximum efficiency.”40 The
linear arch tends to have a rectangular looking curve or a pointed curve (see Fig. 9). This would
greatly explain the evolution of the corbelled shouldered-arch (rectangular curve), the keel arch
and the pointed arch from a structural perspective, as these would be the arches most suitable to
support the weight above them. This occurs because “where an arch is loaded more heavily at its
centre the linear arch will be more sharply pointed, and where the real arch has heavily loaded
haunches the linear arch will be flat-topped.”41
Fig. 9 Linear arches. Rectangular curve, left. Pointed curve, right. (after Gye).
It is not certain whether or not the architects or masons were aware of stability as a result
of the linear arch. If they were, then it was probably knowledge derived from observation, in a
sort of trial and error process, rather than mathematical calculations although the second option
should not be ruled out. However, it seems unlikely that they knew of the linear arch form
39
D.H. Gye, Arches and Domes in Iranian Islamic Buildings: An engineer’s perspective. Iran: Journal of the British
Institute of Persian Studies, no 26 (1988), p. 134
40
Idem.
41
Ibid, p. 136
Abushadi 11
through mathematical calculations because if the linear arch’s shape does not lie within the
actual arch, but is not too far off then the arch must be constructed to be thick in order to
properly support its load without the risk of a collapse. Yet there is no consistent evidence that
shows that the closer arches were in form to the linear arch, the thinner they were as they did not
need to be as thick, which would have saved money on masonry.
Sometimes, what caused the collapse of arches was not an incorrect form, but the stones
or bricks that made up the voussoir of the arch were laid incorrectly, resulting in tension, making
the arch unstable. Arches are strong against compression and sliding, but not against tension42.
It can appear that:
“that tall arches are more stable than flat ones. That is in fact only half true. Although a
flat arch gives rise to large abutment thrusts, it can carry a much larger range of loads
than a tall arch, while still containing the linear arches safely within its thickness. Taking
the flat arch to its ultimate, the lintel orplate-bande, it can in theory withstand any range
of loading whatsoever. However, the thrust at the abutments (and also within the masonry
of the lintel, though this does not matter) can become very high and abutment movement
will therefore be difficult to prevent. The flat arch is, also, simply on account of
geometry, very sensitive to abutment movement; the smallest displacement can result in
substantial sagging”43 (see Fig. 10).
42
43
D.H. Gye, op.cit., p. 131
Ibid, pp. 138-9
Abushadi 12
Sagging would cause tension which could lead to a collapse. There is a correlation between this
and how there is a larger amount of arches in Islamic architecture that have a pointed linear arch
than a rectangular linear arch.
Fig. 10 High rise arch can withstand a smaller load than a low rise arch (after Gye).
It is slightly odd that pointed arches are a lot more common than the rare Islamic
parabolic arches when “for a perfectly uniform load, a rare occurrence in practice, the linear arch
is parabolic”44 (see Fig. 11). However, this might explain why structurally the shouldered and
pointed arch developed into the tri-lobed arch. One can see that the tri-lobed arch rough outline
slightly resembles that of a parabola. As already discussed with the origins of the tri-lobed arch,
it is obvious that structure took a large part in its development, through its place as a squinch in
Iranian Islamic architecture, taking on an important load of the domes it supported.
44
D.H. Gye, op.cit., p. 136
Abushadi 13
Fig. 11 Parabolic linear arch for uniform load (after Gye).
We can see that the shouldered arch developed in a structural manner with the intention
of removing unsupported lintels or to allow for a wider opening without the need for longer
lintels45. Reasons why this might have been desirable include a lack of availability of materials
with which to make suitable lintels, or because the desired width of the opening could not be
achieved with the available lintels46. Gye’s observations showed “that in the corbelled
shouldered-arch form this device can withstand a wide range of loading; it is significant that the
device is used regularly for vaulting in caravanserais, e.g. at Zafaraniyya”47 (see Fig. 10), which
explains why an arch would be preferable to a lintel or wooden beam. Furthermore, in a large
expensive structure, such as a fort, it would save space and hence money to keep the apex height
as low as possible using the shouldered arch without having to import large amounts of wood48.
When structural considerations are taken into account, corbelled shouldered-arches are
preferably used in small openings where the central stone or timber lintel of the arch can be
small and therefore easily obtained. For this reason it is rare to find large arches that are
45
D.H. Gye, op.cit., p. 71
Camilla Edwards & David Edwards, op.cit., p. 71
47
Idem.
48
Idem.
46
Abushadi 14
corbelled shouldered-arches, especially if they are made from brick, even though this is not the
case at Zafaraniyya where the large corbelled shouldered arches are made from brick 49.
It is often common to see tie-beams or tie-bars connecting arches, particularly in
mosques, to decrease the thrust at the abutments50. However, the majority are used in riwaq style
arcades, where the tie-beams are not required51.
Arches are constructed from either brick or stone. Usually the eastern end of the Islamic
Empire uses bricks to construct its arches, while the western end uses stone and the middle
section uses both depending on the nearness of stone quarries.
The general usage of one particular in the construction of arches contributed to which
method of construction is mostly used in which region of the empire. The first method consists
of raising two permanent centerings of reinforced mortar. Then bricks are laid vertically between
the two guides52 (see Fig. 12). It is a method that unsurprisingly is used in the eastern end of the
empire more than the western end. The second method consists of a wooden centering that is
temporarily placed over the jambs. It is then layered with mortar reinforced with small stones.
Then a ring of voussoir is layered around it53 (see Fig. 13). This technique is commonly used
with both bricks and stones.
It has become apparent that within the Islamic Empire arches have evolved through time
and geography. Having started out quite simple as the semicircular arch, it evolved splitting its
centre to form the two-pointed arch which led to the multiple centre arches, the keel arch and the
tri-lobed arch, becoming more complex with the passing of time. Even though, the tri-lobed arch
was likely to have developed as a result of the joint influences of the pointed arch and the
49
Camilla Edwards & David Edwards, op.cit., p. 71
D.H. Gye, op.cit., p. 139
51
Idem.
52
K.A.C. Creswell, Early Muslim Architecture, New York: Hacker Art Books, 1979, vol. 2, p. 61
53
Ibid, p.63
50
Abushadi 15
corbelled shouldered arch. Furthermore, although the origins of the horse-shoe arch in Islamic
architecture are not certain, it is quite possible that it has been influenced by Indian, Christian
Mesopotamian and Romanesque architecture.
Fig. 12 First method of construction (after Creswell).
Fig. 13 Second method of construction (after
Creswell).
Arches have also evolved by regions, where pre-1250, the stone pointed and horse-shoe
arches became popular in the west of Dar al-Islam, whereas the brick pointed, tri-lobed and
corbelled shouldered arch became popular in the east of Dar al-Islam, even though there appears
to be no particular reason for this.
Having looked at the structure of the arches, it has become evident that this evolution in
the form of arches took place not just for aesthetic reasons, but also for structural reasons. Even
though, the evolution of the arch for structural reasons went hand-in-hand with its aesthetic
evolution. However, as mentioned, structurally the arches might have developed in this fashion
as a result of trial and error, where arches close in shape to the linear arch withstood their load
and did not collapse. With the result that arches with a form that lied along the linear arch were
Abushadi 16
the most stable: the pointed arch, the corbelled shouldered arch and a parabolic arch. Which can
also include the keel, tri-lobed and horse-shoe arches as their form generally lies along the
pointed and parabolic linear arches. It is therefore, not surprising that having evolved into these
forms, these arches were the most prominent arches within the Islamic Empire by 1250.
Abushadi 17
References
Behrens-Abouseif, Doris (1989). Islamic architecture in Cairo: An introduction. Cairo, Egypt:
The American University in Cairo Press.
Creswell, K. A. C. (1979). The Muslim architecture of Egypt (Vols. 1-2) (3rd ed.). New York,
NY: Hacker Art Books.
Edwards, Camilla, & Edwards, David (1999). The evolution of the shouldered arch in medieval
Islamic architecture. Architectural History, 42, 68-95. doi:10.2307/j100442
Gye, D.H. (1988). Arches and domes in Iranian Islamic buildings: An engineer’s perspective.
Iran: Journal of the British Institute of Persian Studies, 26, 129-144. doi:10.2307/j101377
Poole, Thomas H. (1913). Arch. Catholic Encyclopaedia. Retrieved from
http://www.catholicliturgy.com/index.cfm/FuseAction/EncyclopediaArticle/Index/15/Sub
Index/149/EncyclopediaIndex/27
Pope, Arthur Upham (1969). Persian Architecture. London, England: Oxford University Press.
Warren, John (1991). Creswell's use of the theory of dating by the acuteness of the pointed
arches in early Muslim architecture. Muqarnas, 8, 59-65. doi:10.2307/j100727